Protection and Robustness: AI systems may be vulnerable to adversarial attacks, where little, imperceptible adjustments for the input may cause the process for making faulty and even hazardous decisions. Making sure the protection and robustness of AI techniques is really a crucial obstacle.